Multi-Step Load Demand Forecasting Using Neural Network

Abstract

The accuracy of load demand forecasting plays a vital role in economic operation and planning in the power sector. Therefore, many techniques and approaches have been proposed in the literature for forecasting. However, there is still an essential need to develop more accurate load forecast method. In this paper, three different strategies of Multi-Step-Ahead Load Forecasting (MSALF), i.e. Direct Strategy (DS), Recursive Strategy (RS) and DirRec Strategy (Direct-Recursive Strategy or DRS) have been used for electricity load demand forecasting by using the Artificial Neural Network (ANN) with Levenberg-Marquardt (LM) training algorithm. The performance evaluation for three different strategies of MSALF has been analysed on two different substations of NE-ISO data sets. Each data sets is analysed for four different cases. The performance of the DRS is better than DS and RS.